INDEX

Explanations

phrases that indicate a conflict of interest or issues pertaining to ethics

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.01

2:0.17

3:0.07

4:0.15

5:0.03

6:0.04

7:0.28

8:0.03

9:0.04

10:0.06

11:0.07

Negative Logits

�

-1.45

ּ

-1.42

Zup

-1.42

 Route

-1.42

aired

-1.39

entanyl

-1.39

�

-1.37

ELS

-1.37

�

-1.36

�

-1.33

POSITIVE LOGITS

 preserving

1.80

 confidentiality

1.77

 autonomy

1.73

 preservation

1.72

 transparency

1.65

 harming

1.64

 moratorium

1.64

 maximizing

1.59

 JUSTICE

1.57

 newsp

1.57

Activations Density 0.001%

phrases that indicate a conflict of interest or issues pertaining to ethics

No Comments

No Known Activations